To maintain a healthy body, the ability of our cells to survive and divide is normally strictly controlled. If any cells manage to escape these restrictions, they may rapidly divide and form tumors, which can lead to cancer.

A protein called AKT can encourage cells to survive and divide, and in healthy cells it is only allowed to be active at specific times. However, in many cancer cells, the genes that make and control AKT activity can be altered by mutations, which can result in AKT being active at the wrong times.

Part of the AKT protein acts as an enzyme called a kinase and adds chemical groups called phosphates to other proteins. The phosphate groups can activate or deactivate these proteins to control cell survival and cell division. However, there are other sections to the AKT protein and it is not clear how they are involved in this protein's activity.

In this study, Vivanco et al. show that AKT has another role in cell survival that does not depend on its kinase. The experiments show that even when the kinase part of the protein is missing, AKT can help cancer cells to survive drug treatments and external conditions that would normally kill them. This role requires another section of the protein called the PH-domain.

There are several chemicals—called inhibitors—that can stop AKT from working properly, and they have the potential to be used to treat some types of cancer. These inhibitors work in different ways: some were able to block the activity of the kinase, but others inhibited AKT by binding to other parts of the protein. Therefore, to develop AKT inhibitors into effective drugs, it will be important to know precisely what role the protein plays in different types of cancers.

The serine–threonine kinase AKT is a non-receptor kinase that plays critical roles in growth and metabolism (Pearce et al., 2010). AKT is aberrantly activated in many human cancers, usually due to mutations in upstream components of the Phosphatidyl-Inositol-3-Kinase (PI3K) pathway. AKT signaling is required for tumor maintenance in certain genetic contexts, as has been shown for some cancer cell lines harboring PI3K mutations or amplification of the HER2 receptor tyrosine kinase (She et al., 2008). Encouraged by these results, several ATP-competitive and allosteric AKT inhibitors have been identified (Engelman, 2009), but optimal strategies for the clinical development of these agents are currently unknown.

To address this question, we compared the activity of the allosteric AKT inhibitor MK2206 (Hirai et al., 2010) with the ATP-competitive AKT inhibitors GSK690693 (Heerding et al., 2008; Rhodes et al., 2008) and GDC0068 (Lin et al., 2013). Each of these AKT inhibitors dose dependently inhibited AKT kinase activity in the non-small-cell lung carcinoma (NSCLC) cell line EBC1 as determined by phosphorylation of three AKT substrates (PRAS40, GSK3β, and BAD) in whole-cell lysates (Figure 1A and Figure 1—figure supplement 1). Consistent with previous reports (Okuzumi et al., 2009), the ATP-competitive AKT inhibitors induced hyper-phosphorylation of AKT at the two regulatory phosphorylation sites threonine 308 and serine 473. When we compared the ability of the AKT inhibitors to induce cell death using a trypan blue exclusion assay, we observed significantly greater cell death induction by the allosteric inhibitor MK2206 compared to the ATP-competitive AKT inhibitors GSK690693 (Figure 1B—source data 1) and GDC0068 (Figure 1C—source data 1) despite less complete inhibition of AKT substrate phosphorylation at these drug concentrations (Figure 1A).

Figure 1 with 9 supplements see all

Download asset Open asset

Figure 1—source data 1

Contains source data for Figure 1 and all accompanying Figure 1—figure supplements.

https://doi.org/10.7554/eLife.03751.004

Download elife-03751-fig1-data1-v2.xlsx

We had selected EBC1 lung cancer cells for our initial studies because they harbor amplification of MET, a receptor tyrosine kinase that mediates critical oncogenic signals through the PI3K/AKT axis (Gherardi et al., 2012). To test whether other MET-amplified human cancer cells show a similar drug response pattern, we extended our experiments to MET-amplified H1993 and H1648 NSCLC cell lines and GTL-16 gastric cancer cells. MK2206 again induced significantly more cell death than GSK690693 (Figure 1—figure supplement 2, Figure 1—source data 1). In cell lines with other PI3K pathway alterations previously associated with ‘AKT-dependence’ (She et al., 2008), we also found that MK2006 induced significantly more cell death than ATP-competitive AKT inhibitors. These cell lines included breast cancer cells with HER2 amplification and/or PIK3CA-mutation (Figure 1D and Figure 1—figure supplement 3, Figure 1—source data 1). In contrast, MK2206 did not induce cell death in non-transformed bronchial epithelial cells (data not shown). MK2206 also did not induce cell death in lung cancer cell lines with activating EGFR mutations (Figure 1—figure supplement 4, Figure 1—source data 1) which have been linked with aberrant PI3K pathway activation (Engelman et al., 2005), suggesting that the mode of PI3K activation is critical in establishing AKT addiction in cancer.

To exclude the possibility that the greater cell death induction by the allosteric AKT inhibitor was due to additional effects on proteins other than AKT, we stably expressed a synthetic AKT mutant in which a residue critical for allosteric inhibitor binding (tryptophan 80) (Green et al., 2008; Calleja et al., 2009; Wu et al., 2010) had been replaced by an alanine. Expression of AKT1 W80A or AKT2 W80A but not wild-type AKT1 or wild-type AKT2 completely rescued MK2206-induced cell death (Figure 1E and Figure 1—figure supplement 5, Figure 1—source data 1). Western blot analysis confirmed that these AKT alleles are biochemically insensitive to MK2206 and prevented the induction of PARP cleavage, a marker of apoptosis (Figure 1—figure supplement 6, Figure 1—source data 1).

Of note, all three inhibitors were able to induce near-complete proliferation arrest (Figure 1—figure supplement 7, Figure 1—source data 1). These results demonstrate that allosteric and ATP-competitive AKT inhibitors differ in their ability to block AKT-mediated survival, and that this survival signal can be mediated by AKT1 or AKT2.

AKT regulates survival through phosphorylation of several protein substrates. We were therefore surprised that ATP-competitive inhibitors induced less cell death than MK2206 despite greater or equal inhibition of AKT substrate phosphorylation (Figure 1A and Figure 1—figure supplement 8). We could not explain these results by differential effects on induction of ErbB3 expression or IGF-IRβ phosphorylation (Figure 1—figure supplement 9) which can accompany AKT inhibition and mediate resistance to AKT inhibitors (Chandarlapaty et al., 2011).

One potential explanation for our observations is that ATP-competitive AKT inhibitors might induce an AKT conformation which can serve as a platform for a kinase-independent survival signal (Figure 2A). If true, ATP-competitive AKT inhibitors might antagonize cell death induction by a variety of cell death stimuli, for example, inhibition of the MET kinase in MET-amplified EBC1 cells. We found this to be the case. Despite inducing a modest amount of cell death as single agent, GSK690693 antagonized the ability of crizotinib to kill EBC1 cells. The allosteric AKT inhibitor MK2206, in contrast, did not compromise the ability of crizotinib to induce cell death (Figure 2B, Figure 2—source data 1).

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

Contains source data for Figure 2 and Figure 2—figure supplement 1.

https://doi.org/10.7554/eLife.03751.015

Download elife-03751-fig2-data1-v2.xlsx

To provide genetic evidence for the kinase-independent AKT function, we took advantage of a synthetic AKT mutant which is catalytically inactive due to substitution of lysine by methionine at position 179 at the ATP-binding site (AKT K179M) (Franke et al., 1995). When we expressed this mutant in MCF7 breast cancer cells, MDA-MB-361 breast cancer cells and EBC1 lung cancer cells, catalytically inactive AKT (Figure 2C/D/E, Figure 2—source data 1) but not wild-type AKT (Figure 2—figure supplement 1, Figure 2—source data 1), protected from MK2206-induced cell death. Compared to the complete cell death protection provided by the drug-resistant W80A-AKT allele (Figure 1E, Figure 1—source data 1), the protection by the K179M-AKT allele was only partial, consistent with the coexistence of kinase-independent and kinase-dependent AKT signals.

We also used an AKT knockdown strategy to minimize the effect of AKT overexpression on the stoichiometry of the AKT signaling complex. For this purpose, we first generated EBC1 cells with doxycycline-inducible human-specific shRNAs for both AKT 1 and AKT 2 and then expressed a kinase-deficient (AKT1-K179M) AKT1 cDNA of murine origin that is partially resistant to knockdown due to mismatches in the mouse sequence. Protein levels of the hairpin-resistant AKT allele after doxycycline induction were similar to the levels of endogenous AKT in parental EBC1 cells (Figure 2F, top, lanes 6 and 8 vs lanes 1 and 3). Similar to our results with pharmacological AKT blockade, catalytically inactive AKT protected from cell death following doxycycline treatment and AKT depletion of these cells (Figure 2F, bottom, Figure 2—source data 1).

We next examined the effects of AKT inhibitors on phosphoinositide (PI) binding. In the absence of growth factors, AKT is inactive because intermolecular interactions between the pleckstrin homology (PH) and kinase domain (KD) keep the kinase in a closed conformation (Calleja et al., 2009). Following growth factor stimulation, AKT shifts to an open conformation and becomes accessible to phosphorylation on threonine 308 and serine 473, a process that requires localization of AKT to the plasma membrane through the interaction with the phospholipid products of PI3K, phosphatidyl-inositol-3,4,5-trisphosphate [PtdIns (3,4,5)P₃], and PtdIns(3,4)P₂. Using a PI pull-down assay, we observed that MK2206 markedly impaired phosphoinositide binding (Figure 3A). This result is consistent with findings from structural studies that allosteric AKT inhibitors lock AKT in a closed conformation with its phospholipid binding site blocked by the kinase domain (Wu et al., 2010). In contrast to our findings with MK2206, incubation with the ATP-competitive inhibitor GSK690693 increased binding of AKT to PtdIns(3,4)P₂ and PtdIns (3,4,5)P₃ and this was associated with increased localization of AKT to the plasma membrane (Figure 3B).

Figure 3 with 4 supplements see all

Download asset Open asset

Figure 3—source data 1

Contains source data for Figure 3 and Figure 3—figure supplement 2.

https://doi.org/10.7554/eLife.03751.018

Download elife-03751-fig3-data1-v2.xlsx

We next explored the structural requirements for the pro-survival signal of catalytically inactive AKT. Similar to GSK690693-bound AKT, the AKT1-K179M (or AKT2-K181M) mutant was hyper-phosphorylated at T308 (Figure 3C and Figure 3—figure supplement 1). This phosphorylation event occurs at the plasma membrane (Stephens et al., 1998) and immunofluorescence in MCF10A and EBC1 cells confirmed that kinase-dead AKT was constitutively localized to the plasma membrane (Figure 3D and Figure 3—figure supplement 2).

To examine whether membrane localization was sufficient for the pro-survival function of AKT1-K179M, we replaced the PH-domain of AKT1-K179M by a membrane-targeting myristoylation signal and expressed this mutant in EBC1 cells. This mutant was unable to inhibit MK2206-induced cell death (Figure 3E, Figure 3—source data 1) despite being constitutively phosphorylated in the presence of MK2206 (Figure 3—figure supplement 3). Lastly, we abolished the ability of the AKT1-K179M mutant to bind phospholipids by introducing a second substitution at an amino acid residue (R86) critical for binding of AKT to PtdIns (3,4,5)P₃ and PtdIns(3,4)P₂ (Thomas et al., 2002). Consistent with a loss of membrane localization, the R86A/K179M double mutant showed loss of AKT phosphorylation at threonine 308 when expressed in HCT116 cells that express no endogenous AKT (Ericson et al., 2010) (Figure 3—figure supplement 4A) and it did not protect from MK2206-induced dephosphorylation in cells which do express endogenous AKT (Figure 3—figure supplement 4B). Abrogation of phosphoinositide binding by the R86A substitution did not impair the pro-survival function of kinase-defective AKT (Figure 3F, Figure 3—source data 1). We observed a similar result with a second mutation (R25C) known to abolish phosphoinositide binding (Thomas et al., 2002) (Figure 3—figure supplement 4C, Figure 3—source data 1).

Mutations in the coding sequence of AKT are rare in human cancers but provide important insights into AKT regulation (Parikh et al., 2012). We wondered whether there was precedence in human cancer for kinase-dead AKT mutations. Our survey of published cancer sequencing data identified five AKT2 kinase domain mutations that occurred in at least two independent samples (G161V/E, R170W, I289M, H355Y, R368C/H) (Figure 4A). We engineered each of these mutations into an AKT2 expression vector and stably expressed the mutant protein in HCT116 cells with targeted deletions of AKT1 and AKT2 (HCT116-DKO). AKT2-G161V, but none of the other four kinase domain mutants, showed complete loss of in vitro AKT substrate phosphorylation, comparable to the synthetic kinase-dead K181M-AKT2 mutant included as control in our experiments (Figure 4B). Whole-cell lysates from G161V-AKT2 expressing cells showed loss of phosphorylation for the endogenous AKT kinase substrates PRAS40 and BAD (Figure 4C). In the phosphoinositide pull-down assay, AKT2-G161V showed altered phosphoinositide binding with acquired preference for PtdIns(4,5)P₂, again similar to the synthetic kinase-dead AKT2 mutant (Figure 4D).

Figure 4

Download asset Open asset

Figure 4—source data 1

Contains source data for Figure 4.

https://doi.org/10.7554/eLife.03751.024

Download elife-03751-fig4-data1-v2.xlsx

Since AKT2-G161V was found in a human melanoma sample, we explored the pro-survival potential of this mutant in immortalized human melanocytes. These cells required 12-O-tetradecanoylphorbol-13-acetate (TPA) for survival in culture, as has previously been reported (Arita et al., 1992), but acquired the ability to survive in TPA-deficient media (RPMI) after stable expression of AKT2-G161V (Figure 4E, Figure 4—source data 1).

The increased PI(4,5)P₂ binding of the kinase-dead G161V-AKT2 mutant (Figure 4D) was reminiscent of the most common somatic AKT mutation in human cancer (E17K-AKT1) which localizes to the plasma membrane due to increased affinity for the constitutive plasma membrane lipid PI(4,5)P₂ (Carpten et al., 2007; Landgraf et al., 2008). This raised the question whether kinase activity might be dispensable for the pro-survival functions of the oncogenic E17K-AKT1 mutant. To explore this question, we compared the effects of E17K-AKT1 and a kinase-defective allele of this mutant (E17K/K179M-AKT1) in EBC1 cells. Compared to cells transduced with wild-type AKT, cells transduced with the E17K-AKT1 mutant showed residual AKT phosphorylation at equimolar concentrations of MK2206 (Figure 4F), consistent with the reported reduced sensitivity of E17K-AKT1 to allosteric ATP kinase inhibitors (Calleja et al., 2009; Wu et al., 2010; Parikh et al., 2012). Expression of E17K-AKT1 provided partial protection from MK2206-induced cell death (Figure 4G, Figure 4—source data 1). Interestingly, the E17K/K179M-AKT1 double mutant provided a similar degree of protection as the E17K-AKT1 single mutant, suggesting that this cancer-associated PH-domain mutant can mediate survival signals independent of its catalytic activity in certain cellular contexts.

Lastly, we identified a human endometrial cancer cell line with a kinase domain mutation in AKT1 (G311D) (Barretina et al., 2012) that failed to induce phosphorylation of endogenous AKT substrates in AKT1/2 DKO HCT116 cells (Figure 5A). While knockdown of endogenous AKT1 or AKT2 resulted in growth inhibition, knockdown of AKT1 resulted in substantially greater cell death induction than knockdown of AKT2 (Figure 5B, Figure 5—source data 1), suggesting a state of addiction to signals provided by the mutant AKT1.

Figure 5

Download asset Open asset

Figure 5—source data 1

Contains source data for Figure 5.

https://doi.org/10.7554/eLife.03751.026

Download elife-03751-fig5-data1-v2.xlsx

The work presented here represents the first example of a kinase-independent function of AKT (Figure 4H). Our findings that catalytically inactivating kinase domain mutations (AKT1-K179M, AKT2-K181M, AKT2-G161V) or ATP-competitive inhibitor binding to wild-type AKT can promote phosphoinositide binding document the critical importance of the kinase domain conformation on AKT PH-domain functions. It also suggests that hyper-phosphorylation of AKT in response to ATP-competitive inhibitors (Okuzumi et al., 2009) may not only be due to impaired phosphatase access (Chan et al., 2011) but also due to increased phosphorylation at the plasma membrane. Furthermore, given that the loss of phosphoinositide binding does not completely abolish the pro-survival function of kinase-deficient AKT, we suspect that the PH-domain promotes cell survival through regulation of interacting protein partners, and that perturbations in the kinase domain can influence PH-domain function by altering the ability of AKT to reach the appropriate sub-cellular compartment (likely those richest in PI(4,5)P2) and/or interact with specific effector proteins. Further studies are required to identify the effectors of catalytically inactive AKT and define its contribution to other AKT-regulated cellular processes.

Our findings have implications for the development of drugs to block AKT in cancer. Previous studies have shown that cancer-associated AKT mutations that favor an open conformation show reduced sensitivity to allosteric AKT inhibitors but retain sensitivity to ATP-competitive inhibitors (Calleja et al., 2009; Wu et al., 2010; Parikh et al., 2012). Here, we show that cancer cells with aberrant AKT activation by common PI3K pathway lesions (HER2, PIK3CA, MET) show the opposite drug sensitivity profile with reduced sensitivity to ATP-competitive AKT inhibitors. The inability of ATP-competitive inhibitors to suppress non-catalytic AKT functions in certain genetic contexts may represent a liability compared to allosteric AKT inhibitors. This concept remains to be tested in ongoing clinical trials.

1. 1. Arita Y
   2. O'Driscoll KR
   3. Weinstein IB

   (1992)

   Growth of human melanocyte cultures supported by 12-O-tetradecanoylphorbol-13-acetate is mediated through protein kinase C activation

   Cancer Research 52:4514–4521.

   • Google Scholar
2. 1. Barretina J
   2. Caponigro G
   3. Stransky N
   4. Venkatesan K
   5. Margolin AA
   6. Kim S
   7. Wilson CJ
   8. Lehár J
   9. Kryukov GV
   10. Sonkin D
   11. Reddy A
   12. Liu M
   13. Murray L
   14. Berger MF
   15. Monahan JE
   16. Morais P
   17. Meltzer J
   18. Korejwa A
   19. Jané-Valbuena J
   20. Mapa FA
   21. Thibault J
   22. Bric-Furlong E
   23. Raman P
   24. Shipway A
   25. Engels IH
   26. Cheng J
   27. Yu GK
   28. Yu J
   29. Aspesi P Jr
   30. de Silva M
   31. Jagtap K
   32. Jones MD
   33. Wang L
   34. Hatton C
   35. Palescandolo E
   36. Gupta S
   37. Mahan S
   38. Sougnez C
   39. Onofrio RC
   40. Liefeld T
   41. MacConaill L
   42. Winckler W
   43. Reich M
   44. Li N
   45. Mesirov JP
   46. Gabriel SB
   47. Getz G
   48. Ardlie K
   49. Chan V
   50. Myer VE
   51. Weber BL
   52. Porter J
   53. Warmuth M
   54. Finan P
   55. Harris JL
   56. Meyerson M
   57. Golub TR
   58. Morrissey MP
   59. Sellers WR
   60. Schlegel R
   61. Garraway LA

   (2012) The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity

   Nature 483:603–607.

   https://doi.org/10.1038/nature11003

   • Google Scholar
3. 1. Calleja V
   2. Laguerre M
   3. Parker PJ
   4. Larijani B

   (2009) Role of a novel PH-kinase domain interface in PKB/Akt regulation: structural mechanism for allosteric inhibition

   PLOS Biology 7:e17.

   https://doi.org/10.1371/journal.pbio.1000017

   • Google Scholar
4. 1. Carpten JD
   2. Faber AL
   3. Horn C
   4. Donoho GP
   5. Briggs SL
   6. Robbins CM
   7. Hostetter G
   8. Boguslawski S
   9. Moses TY
   10. Savage S
   11. Uhlik M
   12. Lin A
   13. Du J
   14. Qian YW
   15. Zeckner DJ
   16. Tucker-Kellogg G
   17. Touchman J
   18. Patel K
   19. Mousses S
   20. Bittner M
   21. Schevitz R
   22. Lai MH
   23. Blanchard KL
   24. Thomas JE

   (2007) A transforming mutation in the pleckstrin homology domain of AKT1 in cancer

   Nature 448:439–444.

   https://doi.org/10.1038/nature05933

   • Google Scholar
5. 1. Chan TO
   2. Zhang J
   3. Rodeck U
   4. Pascal JM
   5. Armen RS
   6. Spring M
   7. Dumitru CD
   8. Myers V
   9. Li X
   10. Cheung JY
   11. Feldman AM

   (2011) Resistance of Akt kinases to dephosphorylation through ATP-dependent conformational plasticity

   Proceedings of the National Academy of Sciences of USA 108:E1120–E1127.

   https://doi.org/10.1073/pnas.1109879108

   • Google Scholar
6. 1. Chandarlapaty S
   2. Sawai A
   3. Scaltriti M
   4. Rodrik-Outmezguine V
   5. Grbovic-Huezo O
   6. Serra V
   7. Majumder PK
   8. Baselga J
   9. Rosen N

   (2011) AKT inhibition relieves feedback suppression of receptor tyrosine kinase expression and activity

   Cancer Cell 19:58–71.

   https://doi.org/10.1016/j.ccr.2010.10.031

   • Google Scholar
7. 1. Engelman JA
   2. Jänne PA
   3. Mermel C
   4. Pearlberg J
   5. Mukohara T
   6. Fleet C
   7. Cichowski K
   8. Johnson BE
   9. Cantley LC

   (2005) ErbB-3 mediates phosphoinositide 3-kinase activity in gefitinib-sensitive non-small cell lung cancer cell lines

   Proceedings of the National Academy of Sciences of USA 102:3788–3793.

   https://doi.org/10.1073/pnas.0409773102

   • Google Scholar
8. 1. Engelman JA

   (2009) Targeting PI3K signalling in cancer: opportunities, challenges and limitations

   Nature Reviews Cancer 9:550–562.

   https://doi.org/10.1038/nrc2664

   • Google Scholar
9. 1. Ericson K
   2. Gan C
   3. Cheong I
   4. Rago C
   5. Samuels Y
   6. Velculescu VE
   7. Kinzler KW
   8. Huso DL
   9. Vogelstein B
   10. Papadopoulos N

   (2010) Genetic inactivation of AKT1, AKT2, and PDPK1 in human colorectal cancer cells clarifies their roles in tumor growth regulation

   Proceedings of the National Academy of Sciences of USA 107:2598–2603.

   https://doi.org/10.1073/pnas.0914018107

   • Google Scholar
10. 1. Franke TF
    2. Yang SI
    3. Chan TO
    4. Datta K
    5. Kazlauskas A
    6. Morrison DK
    7. Kaplan DR
    8. Tsichlis PN

    (1995) The protein kinase encoded by the Akt proto-oncogene is a target of the PDGF-activated phosphatidylinositol 3-kinase

    Cell 81:727–736.

    https://doi.org/10.1016/0092-8674(95)90534-0

    • Google Scholar
11. 1. Gherardi E
    2. Birchmeier W
    3. Birchmeier C
    4. Vande Woude G

    (2012) Targeting MET in cancer: rationale and progress

    Nature Reviews Cancer 12:89–103.

    https://doi.org/10.1038/nrc3205

    • Google Scholar
12. 1. Green CJ
    2. Göransson O
    3. Kular GS
    4. Leslie NR
    5. Gray A
    6. Alessi DR
    7. Sakamoto K
    8. Hundal HS

    (2008) Use of Akt inhibitor and a drug-resistant mutant validates a critical role for protein kinase B/Akt in the insulin-dependent regulation of glucose and system A amino acid uptake

    The Journal of Biological Chemistry 283:27653–27667.

    https://doi.org/10.1074/jbc.M802623200

    • Google Scholar
13. 1. Heerding DA
    2. Rhodes N
    3. Leber JD
    4. Clark TJ
    5. Keenan RM
    6. Lafrance LV
    7. Li M
    8. Safonov IG
    9. Takata DT
    10. Venslavsky JW
    11. Yamashita DS
    12. Choudhry AE
    13. Copeland RA
    14. Lai Z
    15. Schaber MD
    16. Tummino PJ
    17. Strum SL
    18. Wood ER
    19. Duckett DR
    20. Eberwein D
    21. Knick VB
    22. Lansing TJ
    23. McConnell RT
    24. Zhang S
    25. Minthorn EA
    26. Concha NO
    27. Warren GL
    28. Kumar R

    (2008) Identification of 4-(2-(4-amino-1,2,5-oxadiazol-3-yl)-1-ethyl-7-{[(3S)-3-piperidinylmethyl]oxy}-1H- imidazo[4,5-c]pyridin-4-yl)-2-methyl-3-butyn-2-ol (GSK690693), a novel inhibitor of AKT kinase

    Journal of Medicinal Chemistry 51:5663–5679.

    https://doi.org/10.1021/jm8004527

    • Google Scholar
14. 1. Hirai H
    2. Sootome H
    3. Nakatsuru Y
    4. Miyama K
    5. Taguchi S
    6. Tsujioka K
    7. Ueno Y
    8. Hatch H
    9. Majumder PK
    10. Pan BS
    11. Kotani H

    (2010) MK-2206, an allosteric Akt inhibitor, enhances antitumor efficacy by standard chemotherapeutic agents or molecular targeted drugs in vitro and in vivo

    Molecular Cancer Therapeutics 9:1956–1967.

    https://doi.org/10.1158/1535-7163.MCT-09-1012

    • Google Scholar
15. 1. Landgraf KE
    2. Pilling C
    3. Falke JJ

    (2008) Molecular mechanism of an oncogenic mutation that alters membrane targeting: Glu17Lys modifies the PIP lipid specificity of the AKT1 PH domain

    Biochemistry 47:12260–12269.

    https://doi.org/10.1021/bi801683k

    • Google Scholar
16. 1. Lin J
    2. Sampath D
    3. Nannini MA
    4. Lee BB
    5. Degtyarev M
    6. Oeh J
    7. Savage H
    8. Guan Z
    9. Hong R
    10. Kassees R
    11. Lee LB
    12. Risom T
    13. Gross S
    14. Liederer BM
    15. Koeppen H
    16. Skelton NJ
    17. Wallin JJ
    18. Belvin M
    19. Punnoose E
    20. Friedman LS
    21. Lin K

    (2013) Targeting activated Akt with GDC-0068, a novel selective Akt inhibitor that is efficacious in multiple tumor models

    Clinical Cancer Research 19:1760–1772.

    https://doi.org/10.1158/1078-0432.CCR-12-3072

    • Google Scholar
17. 1. Okuzumi T
    2. Fiedler D
    3. Zhang C
    4. Gray DC
    5. Aizenstein B
    6. Hoffman R
    7. Shokat KM

    (2009) Inhibitor hijacking of Akt activation

    Nature Chemical Biology 5:484–493.

    https://doi.org/10.1038/nchembio.183

    • Google Scholar
18. 1. Parikh C
    2. Janakiraman V
    3. Wu WI
    4. Foo CK
    5. Kljavin NM
    6. Chaudhuri S
    7. Stawiski E
    8. Lee B
    9. Lin J
    10. Li H
    11. Lorenzo MN
    12. Yuan W
    13. Guillory J
    14. Jackson M
    15. Rondon J
    16. Franke Y
    17. Bowman KK
    18. Sagolla M
    19. Stinson J
    20. Wu TD
    21. Wu J
    22. Stokoe D
    23. Stern HM
    24. Brandhuber BJ
    25. Lin K
    26. Skelton NJ
    27. Seshagiri S

    (2012) Disruption of PH-kinase domain interactions leads to oncogenic activation of AKT in human cancers

    Proceedings of the National Academy of Sciences of USA 109:19368–19373.

    https://doi.org/10.1073/pnas.1204384109

    • Google Scholar
19. 1. Pearce LR
    2. Komander D
    3. Alessi DR

    (2010) The nuts and bolts of AGC protein kinases

    Nature Reviews Molecular Cell Biology 11:9–22.

    https://doi.org/10.1038/nrm2822

    • Google Scholar
20. 1. Rhodes N
    2. Heerding DA
    3. Duckett DR
    4. Eberwein DJ
    5. Knick VB
    6. Lansing TJ
    7. McConnell RT
    8. Gilmer TM
    9. Zhang SY
    10. Robell K
    11. Kahana JA
    12. Geske RS
    13. Kleymenova EV
    14. Choudhry AE
    15. Lai Z
    16. Leber JD
    17. Minthorn EA
    18. Strum SL
    19. Wood ER
    20. Huang PS
    21. Copeland RA
    22. Kumar R

    (2008) Characterization of an Akt kinase inhibitor with potent pharmacodynamic and antitumor activity

    Cancer Research 68:2366–2374.

    https://doi.org/10.1158/0008-5472.CAN-07-5783

    • Google Scholar
21. 1. She QB
    2. Chandarlapaty S
    3. Ye Q
    4. Lobo J
    5. Haskell KM
    6. Leander KR
    7. DeFeo-Jones D
    8. Huber HE
    9. Rosen N

    (2008) Breast tumor cells with PI3K mutation or HER2 amplification are selectively addicted to Akt signaling

    PLOS ONE 3:e3065.

    https://doi.org/10.1371/journal.pone.0003065

    • Google Scholar
22. 1. Stephens L
    2. Anderson K
    3. Stokoe D
    4. Erdjument-Bromage H
    5. Painter GF
    6. Holmes AB
    7. Gaffney PR
    8. Reese CB
    9. McCormick F
    10. Tempst P
    11. Coadwell J
    12. Hawkins PT

    (1998) Protein kinase B kinases that mediate phosphatidylinositol 3,4,5-trisphosphate-dependent activation of protein kinase B

    Science 279:710–714.

    https://doi.org/10.1126/science.279.5351.710

    • Google Scholar
23. 1. Thomas CC
    2. Deak M
    3. Alessi DR
    4. van Aalten DM

    (2002) High-resolution structure of the pleckstrin homology domain of protein kinase b/akt bound to phosphatidylinositol (3,4,5)-trisphosphate

    Current Biology 12:1256–1262.

    https://doi.org/10.1016/S0960-9822(02)00972-7

    • Google Scholar
24. 1. Wu WI
    2. Voegtli WC
    3. Sturgis HL
    4. Dizon FP
    5. Vigers GP
    6. Brandhuber BJ

    (2010) Crystal structure of human AKT1 with an allosteric inhibitor reveals a new mode of kinase inhibition

    PLOS ONE 5:e12913.

    https://doi.org/10.1371/journal.pone.0012913

    • Google Scholar

Thank you for sending your work entitled and Senior, Thank you for sending your work entitled “A kinase-independent function of AKT promotes cancer cell survival” for consideration at eLife. Your article has been favorably evaluated by Tony Hunter (Senior Editor) and 3 reviewers, one of whom is a member of our Board of Reviewing Editors.

The Reviewing and the other reviewers discussed their comments before we reached this decision, The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

Previous studies have established that AKT provides important survival signals mediated by the phosphorylation of protein substrates, Previous studies have established that AKT provides important survival signals mediated by the phosphorylation of protein substrates. Here the authors demonstrate that AKT also promotes survival by a mechanism that is independent of AKT substrate phosphorylation. The identity of this mechanism is not described. Nevertheless, the conclusions of the paper are novel, interesting, and potentially significant for cancer therapy. However, the conclusions of this study are not fully supported by the data presented.

Major points:

1) One weakness of this study is that it is largely based on over-expression of mutant AKT. Controls using WT AKT are not presented (with the exception of Figure 4G); this important control should be included in each experiment. Moreover, studies of AKT expressed at endogenous levels are essential. Only two studies designed to examine AKT expressed at endogenous levels are presented. The first is a shRNA knock-down/rescue with a shRNA-resistant mutant (Figure 2G). This experiment is difficult to interpret because the mutant AKT is clearly over-expressed and knocked down by the shRNA (despite being stated to be shRNA-resistant). The experiment also lacks controls for off-target actions of the shRNA pools. The second study is an shRNA experiment on endometrial cancer cells (Figure 5B) that lacks controls for off-target effects of the shRNA pools. Additional validation of a kinase-independent function of AKT expressed at endogenous levels is required. For example, can the survival function of kinase-dead AKT be phenocopied by blocking upstream pathways that lead to AKT activation?

2) Figure 2C: The authors argue that the ability of ATP-competitive AKT inhibitors to create complexes with their target(s) that are capable of providing survival signals is revealed by their ability to suppress crizotinib-induced cell death. This would be more clearly validated by a comparison with an “allosteric” inhibitor.

3) The protein localization data are unclear (Figures 3B, D), because the conclusions regarding membrane localization are compromised by the apparent membrane ruffling that causes a higher local density of membrane that can give the false appearance of concentration.

4) Figure 3E: The data obtained with mAKT1-K179M lacking a PH domain are difficult to interpret. Is it targeted to the plasma membrane? Is it substantially resident in other membrane compartments? Could these data indicate that membrane targeting is important, rather than indicating that the survival signals are generated away from membranes, as stated?

5) The analysis is focused on cells with gain-of-function of pathways leading to Akt activation, including MET and PI3K-pathway mutations. Data on survival in response to allosteric vs ATP-competitive AKT inhibitors using cells without this gain-of-function would provide some initial insight into the potential cancer v normal-cell therapeutic window.

6) The HA blot in Figure 3–figure supplement 2B is difficult to interpret.

In response to the reviewers’ excellent suggestions, we have conducted a series of additional experiments that clarified several key points brought up by several reviewers:

To address the role of AKT protein levels in our observations with kinase-dead AKT allels, To address the role of AKT protein levels in our observations with kinase-dead AKT allels, we now expressed wildtype AKT in multiple cell lines (EBC1, MCF7, MDA-MB-361) and show that wild-type AKT is not sufficient to promote cell survival. We also provided more experimental detail regarding our experiment where we knocked down endogenous (human) AKT and replaced it with a murine, kinase dead AKT protein.

Regarding the subcellular localization of kinase-deficient AKT mutants, Regarding the subcellular localization of kinase-deficient AKT mutants, we performed additional confocal microscopy experiments. We again found increased plasma membrane localization of kinase-deficient AKT. Furthermore, by re-introducing wild type or kinase-deficient AKT alleles into AKT1/2 double knockout cells, we show enhanced activation loop phosphorylation (T308) in mutant AKT, a phosphorylation event that occurs at the cytoplasmic membrane.

We now document that the allosteric AKT inhibitor, We now document that the allosteric AKT inhibitor, unlike ATP-competitive AKT inhibitors, does not antagonize the ability of a MET kinase inhibitor to induce cell death in MET-amplified cancer cells. This important control was missing from our original manuscript.

Major points:

1) One weakness of this study is that it is largely based on over-expression of mutant AKT. Controls using WT AKT are not presented (with the exception of Figure 4G); this important control should be included in each experiment.

We thank the reviewers for this suggestion and agree that assessing the effects of ectopically expressed wild type AKT represents an important control, We thank the reviewers for this suggestion and agree that assessing the effects of ectopically expressed wild type AKT represents an important control. To address this question, we now overexpressed wild type AKT1 or wild type AKT2 in several cancer cell lines and repeated key experiments with this additional control. Unlike the W80A-AKT mutants, which cannot bind MK2206, expression of wild type AKT1 (Author response image 1A) or AKT2 (Author response image 1B) does not rescue MDA-MB361 breast cancer cells from MK2206-induced cell death (indicated by red arrows below). We made similar observation in EBC1 lung cancer cells (Author response image 1C). These new data are included in our revised manuscript as Figure 1E, Figure 1–figure supplement 5, Figure 2–figure supplement 1, and Figure 4G (also shown below as Author response image 1A, 1B, 1C, and 1D, respectively).

Author response image 1

Download asset Open asset

Moreover, studies of AKT expressed at endogenous levels are essential. Only two studies designed to examine AKT expressed at endogenous levels are presented. The first is a shRNA knock-down/rescue with a shRNA-resistant mutant (Figure 2G). This experiment is difficult to interpret because the mutant AKT is clearly over-expressed and knocked down by the shRNA (despite being stated to be shRNA-resistant). The experiment also lacks controls for off-target actions of the shRNA pools. The second study is an shRNA experiment on endometrial cancer cells (Figure 5B) that lacks controls for off-target effects of the shRNA pools. Additional validation of a kinase-independent function of AKT expressed at endogenous levels is required. For example, can the survival function of kinase-dead AKT be phenocopied by blocking upstream pathways that lead to AKT activation?

We thank the reviewers for the opportunity to clarify these important points.

The reviewer is correct that the shRNA-“resistant” cDNA is, in fact, only partially resistant to our shRNAs. The targeting sequences used to knock down AKT1 in our experiments (CCCGAGGTGCTGGAGGACA and GGCAAGGTGATCCTGGTGA) are based on human AKT1 and we used murine AKT as a “rescue” allele. As indicated in the Materials and methods section of our revised manuscript, human and murine AKT1 are mismatched only at one or two nucleotides (underlined) relative to each of the targeting sequences (CCTGAGGTGCTGGAGGACA and GGGAAGGTGATTCTGGTGA), likely explaining the incomplete rescue. Nonetheless, despite the effects of the shRNA-hairpins on the murine mRNA, only cells expressing the murine AKT1 cDNA show residual AKT1 following doxycycline treatment. [Please compare, in Author response image 2 below, lanes 6 and 8 (green arrows) to lanes 2 and 4 (red arrows).] Of note, levels of residual K179M-mAKT1 (lanes 6 and 8, green arrows) are roughly equivalent to the endogenous AKT levels seen in the absence of doxycycline (lanes 1 and 3, blue arrows), an effect not seen with AKT2 expression. In our revised manuscript, we have revised the figure and text to explain this point more clearly.

Author response image 2

Download asset Open asset

We also examined the question whether the survival function of kinase-dead AKT can be phenocopied by blocking upstream pathways that lead to AKT activation. We found that the ATP-competitive AKT inhibitor GSK690693 does not promote cell survival if we block all AKT activating signals through pharmacological PI3-kinase inhibition with the PI3K inhibitor BKM-120 (Author response image 3). These data suggests that even the kinase-independent functions of AKT require a PI3K input. We are therefore unable to experimentally phenocopy the pro-survival functions of kinase-deficient AKT by blockade of upstream signals.

Author response image 3

Download asset Open asset

2) Figure 2C: The authors argue that the ability of ATP-competitive AKT inhibitors to create complexes with their target(s) that are capable of providing survival signals is revealed by their ability to suppress crizotinib-induced cell death. This would be more clearly validated by a comparison with an “allosteric” inhibitor.

We apologize for not including these data in our prior submission. In our revised manuscript, we are showing the full experiment which included the allosteric AKT inhibitor MK2206 and the crizotinib/MK2206 combination. The results show that the allosteric AKT inhibitor MK2206, unlike the ATP-competitive inhibitor GSK690693, does not protect from crizotinib-induced cell death (Author response image 4), consistent with our model. These data are included as Figure 2B in the revised manuscript.

Author response image 4

Download asset Open asset

3) The protein localization data are unclear (Figures 3B, D), because the conclusions regarding membrane localization are compromised by the apparent membrane ruffling that causes a higher local density of membrane that can give the false appearance of concentration.

We thank the reviewers for the opportunity to clarify this important point. In our original manuscript, we showed increased phosphorylation of the kinase-deficient AKT1 mutant (K179M) at Serine 308 (Author response image 5A). Other investigators have shown that phosphorylation of AKT at T308 occurs at the plasma membrane (Stephens et al., Science 1998, PMID: 9445477), and we believe that this is consistent with our previously provided immunofluorescence data (Author response image 5B).

In response to the reviewers, we have now also examined the cellular localization of kinase deficient AKT2 (K181M). As shown in confocal images of wild type vs. mutant AKT2 in stably-transduced EBC1 cells (Author response image 5C, left column), kinase-deficient AKT2 more robustly colocalizes with the plasma membrane marker Na+/K+ ATPase than its wild type counterpart. Furthermore, following the reviewer’s suggestion, we co-stained these cells with phalloidin which binds F-actin and marks membrane ruffles. We found increased membrane localization of mutant AKT2 even in the absence of membrane ruffles (red arrows in Author response image 5C, right column). These data are included in Figure 3–figure supplement 2 of our revised manuscript.

Author response image 5

Download asset Open asset

We thus conclude that kinase-deficient AKT1 and AKT2 show increased localization to the plasma membraneOur results with engineered and naturally occurring kinase-dead AKT mutants are reminiscent of the findings from Kevan Shokat’s group at UCSFwhich documented relocalization of AKT to the plasma membrane in response to ATP-competitive Akt kinase inhibitors (Okuzumi et al., 2009 We thus conclude that kinase-deficient AKT1 and AKT2 show increased localization to the plasma membrane. Our results with engineered and naturally occurring kinase-dead AKT mutants are reminiscent of the findings from Kevan Shokat’s group at UCSF, which documented relocalization of AKT to the plasma membrane in response to ATP-competitive Akt kinase inhibitors (Okuzumi et al., Nature Chemical Biology, 2009, PMID: 19465931). In contrast, the allosteric AKT inhibitor MK2206 does not promote membrane relocalization of AKT.

4) Figure 3E: The data obtained with my-Akt1-K179M lacking a PH domain are difficult to interpret. Is it targeted to the plasma membrane? Is it substantially resident in other membrane compartments? Could these data indicate that membrane targeting is important, rather than indicating that the survival signals are generated away from membranes, as stated?

We completely agree with the reviewers’ interpretation of the data and have rephrased our conclusions to express this more clearly. The non-catalytic function of AKT requires the presence of the PH domain, as demonstrated by our data with the Δ PH-Myr-Akt1-K179M mutant. Together with the results of our experiments using the R86C and R25C mutants, this data suggest that the kinase-independent function of AKT might require targeting to specific membranes or membrane subdomains, likely those richest in PI(4,5)P2, to which the mutants show enhanced binding.

5) The analysis is focused on cells with gain-of-function of pathways leading to Akt activation, including MET and PI3K-pathway mutations. Data on survival in response to allosteric vs ATP-competitive AKT inhibitors using cells without this gain-of-function would provide some initial insight into the potential cancer v normal-cell therapeutic window.

We thank the reviewers for this insightful suggestion, We thank the reviewers for this insightful suggestion.

To address the issue, To address the issue, we have treated BEAS-2B immortalized (non-transformed) bronchial epithelial cells with MK2206, and found that it could not induce any appreciable amount of cell death (Author response image 6).

We have also treated three NSCLC cell lines carrying endogenous activating EGFR mutations with AKT inhibitors (also shown in Author response image 6), We have also treated three NSCLC cell lines carrying endogenous activating EGFR mutations with AKT inhibitors (also shown in Author response image 6). While we have found that amplification of MET and activating PIK3CA-mutations are associated with sensitivity to AKT allosteric inhibitors, we find that the EGFR-mutant lung cancer cell lines are completely refractory to cell death induction by either MK2206 or GSK690693. Given that these EGFR mutations are capable of activating AKT (PMID: 15731348), these data suggest that the mechanism of PI3K activation (e.g. EGFR activation vs MET activation) is likely critical in establishing AKT dependence. These data have been added to the revised manuscript as Figure 1–figure supplement 4.

Author response image 6

Download asset Open asset

6) The HA blot in Figure 3–figure supplement 2B is difficult to interpret.

We apologize for the poor quality of this blot. We have now performed additional western blot analysis with an AKT1-specific antibody to quantify AKT1 expression in the transduced sublines, relative to the parental controls. The results show increased AKT1 expression in cells transduced with wild type or mutant AKT1 cells (see image quantification below each lane in Author response image 7). Of note, it also again shows that the kinase-dead K179M-AKT1 mutant exhibits increased phosphorylation at threonine 308 (lanes 5) compared to wild type AKt1 (lane 3) and that this hyperphosphorylation is lost when PIP3-binding is disrupted by the second R86A mutation (lane 7). These new data have been included as Figure 3–figure supplement 4B of the revised manuscript.

Author response image 7

Download asset Open asset

Author details

1. Igor Vivanco

   Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, United States

   Present address

   Male Urological Cancer Research Centre, Division of Cancer Therapeutics, Institute of Cancer Research, Sutton, United Kingdom

   Contribution

   IV, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Zhi C Chen

   For correspondence

   igor.vivanco@icr.ac.uk

   Competing interests

   The authors declare that no competing interests exist.

2. Zhi C Chen

   Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, United States

   Contribution

   ZCC, Acquisition of data, Analysis and interpretation of data

   Contributed equally with

   Igor Vivanco

   Competing interests

   The authors declare that no competing interests exist.

3. Barbara Tanos

   Cell Biology Program, Memorial Sloan-Kettering Cancer Center, New York, United States

   Contribution

   BT, Conception and design, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

4. Barbara Oldrini

   Seve Ballesteros Foundation Brain Tumor Group, Spanish National Cancer Research Centre, Madrid, Spain

   Contribution

   BO, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

5. Wan-Ying Hsieh

   Department of Pharmacology, Weill-Cornell Graduate School of Biomedical Sciences, New York, United States

   Contribution

   W-YH, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

6. Nicolas Yannuzzi

   Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, United States

   Contribution

   NY, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

7. Carl Campos

   Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, United States

   Contribution

   CC, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

8. Ingo K Mellinghoff

   1. Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, United States
   2. Department of Pharmacology, Weill-Cornell Graduate School of Biomedical Sciences, New York, United States
   3. Department of Neurology, Memorial Sloan-Kettering Cancer, New York, United States

   Contribution

   IKM, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   mellingi@mskcc.org

   Competing interests

   The authors declare that no competing interests exist.

• 4,465

  Page views

• 686

  Downloads

• 61

  Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.